Benchmark result
Qwen3.8-27B-YMQ-M-TI on 2× NVIDIA GeForce RTX 5060 — 35.0 tok/s
Measured with llama.cpp b11253-86ea01d05 on October 1, 2026.
Support llm-bench.io
llm-bench.io is free. If it helped you, I'd appreciate your support to cover the infrastructure costs.
What the model built
Open full screen →The coding scenario asks for a playable game in a single HTML file. This is exactly what the model returned, unedited; the frame only adds a line that tells this page how tall it is.
How this run compares
6th fastest of 6 runs of this model on 2× NVIDIA GeForce RTX 5060 · Multi-GPU · median 35.9 tok/s.
sort
- 136.9tok/s87.2llama.cpp
- 236.8tok/s83.5llama.cpp
- 336.2tok/s82.4llama.cpp
- 435.7tok/s88.1llama.cpp
- 535.3tok/s81.7llama.cpp
- 635.0tok/s84.8llama.cppthis run
Reproduce this run
toolllama.cpp b11253-86ea01d05
modelQwen3.8-27B-YMQ-M-TI
context81,920
thinkingon
samplingtemperature 1 · top_p 0.949999988079071 · top_k 20 · min_p 0.05000000074505806 · presence_penalty 0 · repeat_penalty 1.0499999523162842
clientv0.4.85+101
Set those in llama.cpp, then:
llm-benchmark benchmark --model "Qwen3.8-27B-YMQ-M-TI" --tool "llama.cpp"The client prompts for context length and thinking mode, and for the KV cache dtype on Unsloth Studio. Sampling is left at the model default — the values above are what the tool reported using, not overrides the client sent.